{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:6KNMSIJZ73JYVQFANB2UVDNVRE","short_pith_number":"pith:6KNMSIJZ","canonical_record":{"source":{"id":"2506.01266","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-02T02:36:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e0c307fd87a1fae0456c873f58fce15a8b53b0ff646b4c2fd3329b9f130ab989","abstract_canon_sha256":"9426612e5eaaaa88c3dd8637a6cfc78cd288f4c519830e45d4cfa24329e41594"},"schema_version":"1.0"},"canonical_sha256":"f29ac92139fed38ac0a068754a8db589058b92e49e8739318b408cce541efefb","source":{"kind":"arxiv","id":"2506.01266","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.01266","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"arxiv_version","alias_value":"2506.01266v1","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.01266","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"pith_short_12","alias_value":"6KNMSIJZ73JY","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"pith_short_16","alias_value":"6KNMSIJZ73JYVQFA","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"pith_short_8","alias_value":"6KNMSIJZ","created_at":"2026-07-05T11:14:10Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:6KNMSIJZ73JYVQFANB2UVDNVRE","target":"record","payload":{"canonical_record":{"source":{"id":"2506.01266","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-02T02:36:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e0c307fd87a1fae0456c873f58fce15a8b53b0ff646b4c2fd3329b9f130ab989","abstract_canon_sha256":"9426612e5eaaaa88c3dd8637a6cfc78cd288f4c519830e45d4cfa24329e41594"},"schema_version":"1.0"},"canonical_sha256":"f29ac92139fed38ac0a068754a8db589058b92e49e8739318b408cce541efefb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:10.928565Z","signature_b64":"hIDdVmRV2lMN72DL6lvX/FJtMnRX5rNSihco22ykti1NbB4vr0JdCQ/m9S4cxbBou0EO6GeUQZg9A8ZCpFMPBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f29ac92139fed38ac0a068754a8db589058b92e49e8739318b408cce541efefb","last_reissued_at":"2026-07-05T11:14:10.927930Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:10.927930Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.01266","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:14:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5X3uqJLVk8+BcTAD/1DPS/PUutbEt4xw5eH8GTM1Tj6q0dOnPlrDqOUzXey0DfGa9whf4zmF9Gv5znVztv61Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T06:38:25.481961Z"},"content_sha256":"301c7158da4019b65d4e8cc95cef3a683cb88c454b9927a8ea574e7aa44ab61a","schema_version":"1.0","event_id":"sha256:301c7158da4019b65d4e8cc95cef3a683cb88c454b9927a8ea574e7aa44ab61a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:6KNMSIJZ73JYVQFANB2UVDNVRE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Detoxification of Large Language Models through Output-layer Fusion with a Calibration Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Guoqing Jin, Mingjie Deng, Yan Song, Yuanhe Tian","submitted_at":"2025-06-02T02:36:32Z","abstract_excerpt":"Existing approaches for Large language model (LLM) detoxification generally rely on training on large-scale non-toxic or human-annotated preference data, designing prompts to instruct the LLM to generate safe content, or modifying the model parameters to remove toxic information, which are computationally expensive, lack robustness, and often compromise LLMs' fluency and contextual understanding. In this paper, we propose a simple yet effective approach for LLM detoxification, which leverages a compact, pre-trained calibration model that guides the detoxification process of a target LLM via a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.01266","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2506.01266/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:14:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RFEATz1kcebxmWaMOBhprIcmxQc6MDmlSLDvBXXgwIs06BRVHIdgWyeOM/td0QmowvcKtjdEBlW6t77Z5bg5Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T06:38:25.482587Z"},"content_sha256":"b89ccba20a147b9c8de12d7c16e3d5564c20b74e4b582e5f5bb3b31e617c264a","schema_version":"1.0","event_id":"sha256:b89ccba20a147b9c8de12d7c16e3d5564c20b74e4b582e5f5bb3b31e617c264a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6KNMSIJZ73JYVQFANB2UVDNVRE/bundle.json","state_url":"https://pith.science/pith/6KNMSIJZ73JYVQFANB2UVDNVRE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6KNMSIJZ73JYVQFANB2UVDNVRE/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-09T06:38:25Z","links":{"resolver":"https://pith.science/pith/6KNMSIJZ73JYVQFANB2UVDNVRE","bundle":"https://pith.science/pith/6KNMSIJZ73JYVQFANB2UVDNVRE/bundle.json","state":"https://pith.science/pith/6KNMSIJZ73JYVQFANB2UVDNVRE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6KNMSIJZ73JYVQFANB2UVDNVRE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:6KNMSIJZ73JYVQFANB2UVDNVRE","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9426612e5eaaaa88c3dd8637a6cfc78cd288f4c519830e45d4cfa24329e41594","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-02T02:36:32Z","title_canon_sha256":"e0c307fd87a1fae0456c873f58fce15a8b53b0ff646b4c2fd3329b9f130ab989"},"schema_version":"1.0","source":{"id":"2506.01266","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.01266","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"arxiv_version","alias_value":"2506.01266v1","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.01266","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"pith_short_12","alias_value":"6KNMSIJZ73JY","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"pith_short_16","alias_value":"6KNMSIJZ73JYVQFA","created_at":"2026-07-05T11:14:10Z"},{"alias_kind":"pith_short_8","alias_value":"6KNMSIJZ","created_at":"2026-07-05T11:14:10Z"}],"graph_snapshots":[{"event_id":"sha256:b89ccba20a147b9c8de12d7c16e3d5564c20b74e4b582e5f5bb3b31e617c264a","target":"graph","created_at":"2026-07-05T11:14:10Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2506.01266/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Existing approaches for Large language model (LLM) detoxification generally rely on training on large-scale non-toxic or human-annotated preference data, designing prompts to instruct the LLM to generate safe content, or modifying the model parameters to remove toxic information, which are computationally expensive, lack robustness, and often compromise LLMs' fluency and contextual understanding. In this paper, we propose a simple yet effective approach for LLM detoxification, which leverages a compact, pre-trained calibration model that guides the detoxification process of a target LLM via a ","authors_text":"Guoqing Jin, Mingjie Deng, Yan Song, Yuanhe Tian","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-02T02:36:32Z","title":"Detoxification of Large Language Models through Output-layer Fusion with a Calibration Model"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.01266","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:301c7158da4019b65d4e8cc95cef3a683cb88c454b9927a8ea574e7aa44ab61a","target":"record","created_at":"2026-07-05T11:14:10Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9426612e5eaaaa88c3dd8637a6cfc78cd288f4c519830e45d4cfa24329e41594","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-02T02:36:32Z","title_canon_sha256":"e0c307fd87a1fae0456c873f58fce15a8b53b0ff646b4c2fd3329b9f130ab989"},"schema_version":"1.0","source":{"id":"2506.01266","kind":"arxiv","version":1}},"canonical_sha256":"f29ac92139fed38ac0a068754a8db589058b92e49e8739318b408cce541efefb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f29ac92139fed38ac0a068754a8db589058b92e49e8739318b408cce541efefb","first_computed_at":"2026-07-05T11:14:10.927930Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:14:10.927930Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hIDdVmRV2lMN72DL6lvX/FJtMnRX5rNSihco22ykti1NbB4vr0JdCQ/m9S4cxbBou0EO6GeUQZg9A8ZCpFMPBg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:14:10.928565Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.01266","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:301c7158da4019b65d4e8cc95cef3a683cb88c454b9927a8ea574e7aa44ab61a","sha256:b89ccba20a147b9c8de12d7c16e3d5564c20b74e4b582e5f5bb3b31e617c264a"],"state_sha256":"6454dcac99e0cd7b84696eaa712c2245ffc1480d94d4cd594dfb762affb0f011"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rdy3sMvK9RTEjcFKKpezefKDmhqrZVu10leZTRW0p4xRvcw25dRU6AZ9mn79g/X9ohl51caHoBvITPdwsxZyDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T06:38:25.490492Z","bundle_sha256":"fa8e85dcdb80e4c278330c099c95aea48ce1445dac51bbdf980a4c9aa998e022"}}